DESCRIPTION: Functional magnetic resonance imaging (fMRI) has enormous potential for both scientific and clinical studies of human brain function. For diagnostic radiology, fMRI's ability to localize functional areas non-invasively promise to make it an especially important tool in presurgical planning for epilepsy and brain tumors. To make this technology practical, however, it must be more reliable than currently available. To accomplish this, fMRI must have high spatial and temporal resolution, good paradigm control, appropriate statistica analysis, and rapid data processing. Another important practical consideration is that fMRI needs to be able to yield results even under less than ideal subject performance conditions. In clinical studies, patients are typically no extensively trained in their tasks and tend to be highly variable in their performance. Scan reliability and spatial resolution can be effectively improved by explicitly separating MR signal changes associated with task performance, away from signal changes attributable to other physiological, behavioral, electrical, or mechanical variables. The Principal Investigator (PI) has developed real-time paradigm control software and real-time image analysis software for fMRI that generalize this strategy for improving scan quality by allowing many different behavioral and physiological variables to be integrate simultaneously. An important aspect of this approach is that all the analysis is performed in close to real-time, so that the complete results are available during the scan or within a minute after scan completion. This immediate feedback provides quality assurance that greatly improves scan reliability. Fo clinical scans, such timely results are critical if fMRI is to be used to tailor treatment to individual patients. The current application seeks support to apply and extend this work to identif optimal software strategies for improving the effective sensitivity and resolution of clinical and research fMRI. This project tests the hypothesis that physiological and behavioral data improves task-specific signal-to-noise ratio and spatial resolution. Subjective behavioral responses will be tested a probes of cognitive brain function. Practical strategies for obtaining comprehensive fMRI results in close to real-time will be investigated. Integrating the strategies addressed in this study will make fMRI a more sensitive and reliable brain imaging tool, thereby extending its capabilities for a wide range of both scientific and clinical applications.